Mind Your Heart-II: Protocol for a Behavioral Randomized Controlled Trial of Mindfulness Training to Promote Self-Care in Patients with Comorbid Heart Failure and Cognitive Impairment

Abstract

Background:

Heart failure (HF) self-care is a robust predictor of prognosis in HF patients. Cognitive impairment is a common comorbidity in HF patients and constitutes a major challenge to HF self-care. Mindfulness training (MT) has been shown to improve cognitive function and interoception, two components essential to promoting effective HF self-care.

Objectives:

The aims of the Mind Your Heart-II (MYH-II) study are to investigate the effects of MT on HF self-care via changes in cognitive function and interoception in patients with comorbid HF and cognitive impairment, and to study the process by which MT can improve cognitive function via vagal control. We hypothesize that MT will improve cognitive function, interoception, and vagal control, resulting in enhanced HF self-care, compared to control participants.

Methods:

MYH-II is a mechanistic parallel phase II behavioral randomized controlled trial. We will enroll 176 English or Spanish-speaking patients with comorbid chronic HF and mild cognitive impairment. Participants will be randomized to either: (1) 8-week phone-delivered MT + Enhanced Usual Care (EUC), or (2) EUC alone. Participants will complete baseline, end-of-treatment (3 months), and follow-up (9 months) assessments. The primary outcome is cognitive function (NIH Toolbox Fluid Cognition Composite Score). Additional key outcomes include: interoception (heartbeat tracking task, Multidimensional Assessment of Interoceptive Awareness), HF self-care (Self-Care of Heart Failure Index v7.2), and vagal control (high-frequency heart rate variability).

Implications:

If study hypotheses are confirmed, phone-based MT may be a key tool for improving HF self-care, and possibly clinical outcomes, in HF patients with comorbid cognitive impairment.

Introduction

As the most common diagnosis at hospital discharge among older adults,¹ heart failure (HF) is a significant public health concern. Rates of 6-month hospital readmissions and 5-year mortality reach 50%,^2–6 and treatment costs continue to soar.¹ To improve clinical outcomes, it is critical that patients with HF perform adequate self-care, i.e., adhere to pharmacological and lifestyle recommendations and monitor their symptoms to ensure they receive medical attention in a timely fashion.^7–11 The situation specific theory of HF self-care identifies maintenance (i.e., treatment adherence), symptom perception (i.e., monitoring, detection, and interpretation of symptoms), and management (i.e., symptom response) self-care components.¹² Without the ability to perceive symptoms, patients with HF cannot be expected to engage in self-care management with an appropriate symptom response. Importantly, one critical barrier to each of the aspects of HF self-care is cognitive impairment, which affects up to 2/3 of HF patients.^13–16

Mindfulness training (MT) is an intervention designed to promote mindfulness, the “non-judgmental, sustained, moment-to-moment awareness of physical sensations, affective states, and thoughts.”¹⁷ There are preliminary indications that MT can improve cognitive function in older adults, and thus, may facilitate self-care in this population. The authors have conducted two pilot studies supporting the potential benefits of MT on cognitive function in older adults at-risk for dementia^18,19 and patients with HF.²⁰ These studies have provided preliminary support for both class- and phone-delivered MT improving cognitive function (assessed using standard neurocognitive tests), including executive function, memory, and attention. In addition, recent systematic reviews have shown that MT can improve cognitive function – particularly executive function and working memory – in older adults with and without mild cognitive impairment and/or dementia, with small-to-medium effect sizes.^21,22 Despite these promising findings, the validity of the current evidence is threatened by high heterogeneity, inconsistencies in intervention strategies and outcome measures, and the small sample size of many studies.

In addition to promoting cognitive function, there is preliminary evidence that MT may improve interoception, i.e., the ability to detect and respond to physical sensations originating within the body (e.g., heartbeat, breath).²³ Although relatively understudied in this population, interoception could be a key factor in facilitating HF self-care (e.g., symptom perception) as it is associated with the ability to detect HF symptoms.²⁴ MT also elicits the “relaxation response,” which is characterized by a decrease in sympathetic nervous system activity and an increase in autonomic vagal control.^25,26 Improved vagal control has been shown to have positive effects on cognitive function,²⁷ and thus is an important physiological mechanism by which MT could positively affect cognitive function.

The current phase II behavioral randomized clinical trial (RCT) will be the first large-scale investigation of the effects of MT on cognitive function in HF patients with comorbid cognitive impairment, and the first to assess whether changes in cognitive function are associated with improved HF self-care. A composite cognitive function score including executive function, attention, memory, and processing speed domains is the primary outcome of this study. This RCT will additionally study interoception as an alternative pathway whereby MT may improve HF self-care. Lastly, we will investigate vagal control as a possible physiological mediator of the relationship between MT and improvements in cognitive function. These proposed pathways between MT, cognitive function, interoception, vagal control, and HF self-care are summarized in Figure 1.

Figure 1.

Conceptual Model Illustrating Proposed Pathways

This RCT will specifically investigate: 1) the role of MT in improving cognitive function and HF self-care in patients with comorbid HF and cognitive impairment, 2) the role of MT in improving interoception and HF self-care, and 3) the mechanistic pathway linking MT, vagal control, and cognitive function. We hypothesize that: a) MT plus enhanced usual care (MT+EUC), vs. EUC, will improve cognitive function and interoception at end-of-treatment, and changes in these factors will be associated with improvements in HF self-care at follow-up, and b) MT+EUC, vs. EUC, will improve vagal control at end-of-treatment, and these changes will be associated with improvements in cognitive function at follow-up.

Methods

Study Design

The Mind Your Heart-II (MYH-II) study is a mechanistic, parallel, phase II behavioral RCT (ClinicalTrials.gov ID:NCT05431192).

Setting

MYH-II will be conducted within the Lifespan healthcare system, an affiliate of the Alpert Medical School of Brown University. Recruitment hospitals will include the Rhode Island Hospital and The Miriam Hospital (Providence, RI) and the Newport Hospital (Newport, RI). Study assessments will be conducted at the Lifespan Cardiovascular Institute.

Study Procedures Overview

MYH-II includes an 8-week intervention period and baseline (Visit 1), end-of-treatment (3 months; Visit 2), and follow-up (9 months; Visit 3) assessments (Figure 2). The mindfulness intervention will be conducted via phone by trained mindfulness instructors. All assessments will be completed in-person by trained study personnel. Visit 1 will include eligibility screening and baseline assessment of study outcomes. Visits 2 and 3 will include end-of-treatment and follow-up assessments, respectively. Study procedures are detailed below and have been approved by The Miriam Hospital’s Institutional Review Board (IRB).

Figure 2.

Overview of Study Procedures

Population

Participants will be adults recruited from HF clinics at the designated hospitals or from the Providence community. Inclusion criteria include: 1) Age >18 years, 2) Mild cognitive impairment (i.e., MoCA 15–26), 3) Documented diagnosis of HF in electronic medical record (EMR), 4) Access to a telephone, and 5) Ability to understand and speak English or Spanish. Participants will be excluded if any of the following conditions are met: 1) Unwillingness/inability to provide informed consent, 2) Reversible causes of HF (e.g., myocarditis), 3) Severe hearing impairment not allowing phone delivery, 4) Current suicidal ideation or plan, 5) Current (≥ once/month) mind-body practice, 6) Planning to move out of the area during the study period, 7) Severe cognitive impairment (MoCA <15), 8) New York Heart Association²⁸ class IV HF or clinically unstable, 9) Ongoing severe psychiatric or neurologic conditions documented in EMR, or 10) Current enrollment in any other study.

Recruitment and Screening

Recruitment strategies include EMR review; flyers posted at community centers, libraries, and supermarkets; advertisements in community newspapers; in-person referrals; and on-line resources. Potentially eligible participants identified via EMR will be mailed a study invitation and information packet, followed by a phone call from the project director (PD). Individuals recruited from the community will be directed to call the PD to express interest. Thus, all interested participants will complete a preliminary phone screening (using a scripted screening instrument) with the PD. The PD will then schedule an in-person screening visit for qualifying individuals.

In-Person Screening – Visit 1

At Visit 1 participants will provide informed consent and undergo assessment of remaining eligibility criteria. To screen for cognitive impairment, participants will complete the Montreal Cognitive Assessment (MoCA),²⁹ a 30-item valid and reliable instrument widely used in the HF literature with a possible score of 0–30. A score of 27–30 = ‘no to light CI’; 18–26 = ‘mild CI’; 10–17 = ‘moderate CI’; <10 = ‘severe CI’. Participants with MoCA score ≤26 (mild CI) will be included in the study. Patients with MoCA score <15 will be excluded because this intervention has not been tested in populations with more severe cognitive impairment. If scores indicating significant cognitive impairment are detected, the participant’s primary care provider will be notified.

Suicidal ideation or plan will be assessed using the Symptom Driven Diagnostic System for Primary Care, a validated screening instrument.^30,31 Individuals endorsing ≥1 item will be excluded from the study and will complete a suicide risk assessment following a standard protocol (see Safety – Emotional Discomfort). Eligible individuals will then complete study baseline assessments and randomization before concluding the visit.

Randomization

We will use block randomization stratified by gender and age, generated in “R”³² by the study statistician, to allocate participants to MT+EUC or EUC. The allocation table will be uploaded to an Access database. After completion of baseline assessments, the research assistant (RA) will randomly assign study participants (1:1 ratio) by clicking on the “randomize” icon and immediately inform the participant of their assigned treatment condition.

Study Interventions

Mindfulness Training (MT)

The MT protocol maintains the basic components of Mindfulness-Based Stress Reduction (MBSR), streamlined to meet the needs of individuals with chronic conditions and maximize adherence. Traditional, in-person MT programs (e.g., 8 weekly classes lasting 2.5 hours, one full-day workshop, and intensive individual practice)³³ are not suitable for older individuals with chronic medical conditions such as HF. Thus, we have developed a phone-delivered MT consisting of eight weekly 30-minute phone-based MT sessions and associated individual home practice for 20 minutes daily with the guidance of a digital mindfulness recording. This adapted MT protocol has shown high feasibility and acceptability in previous studies of participants with a range of chronic conditions, including HF.^34–37

MT topics include (Table 1): (1) Awareness of breath, a technique in which trainees learn to attend to the sensations associated with breathing; (2) Body scan, a technique based on the cultivation of attention to bodily sensations that would normally go unnoticed. Later, participants are trained (3) to direct their attention to simple activities of daily life; (4) to become aware of their own thoughts and emotions; and (5) to practice “open awareness” – a technique whereby participants are invited to direct their attention to any event arising in their field of experience at a given moment (e.g., physical sensation, sound, emotion, thought). MT treatment fidelity will be monitored per Treatment Fidelity Workgroup guidelines.³⁸

Table 1.

Session by Session Overview of the Mindfulness Training Intervention Components

	Session Number
	1	2	3	4	5	6	7	8
Awareness of Breath	X	X	X	X	X	X	X	X
Body Scan			X	X	X	X	X	X
Awareness of Daily Experience Exercise		X
Informal Practice Exercise				X
Awareness of Sounds Exercise					X			X
Awareness of Emotions Exercise						X		X
Awareness of Thoughts Exercise							X	X
Open Awareness								X

MT Instructors.

The MT instructors will be graduates of the teachers’ training program at the University of Massachusetts Center for Mindfulness with ≥5 years teaching experience and previous experience with studies involving phone-delivered MT. To ensure consistency of delivery, each participant will be trained by the same instructor throughout the intervention. Instructors will receive one 2-hour training session (review of MT protocol; fidelity procedures).

Enhanced Usual Care (EUC)

The American Heart Association recommends that all patients with HF receive education to facilitate HF self-care; thus, participants in both conditions will receive a printed version of the “Healthier Living with Heart Failure” guide³⁹ (in addition to the usual care offered by their health care providers) – a publicly available educational resource. This guide includes seven chapters addressing topics such as understanding HF, making healthy lifestyle changes, managing HF symptoms, taking medications, and living well with HF. Educational content will be mailed weekly to all participants during the intervention period. To ensure that no unwarranted “care” is offered besides that of healthcare providers (e.g., informal support from study staff), we will limit unnecessary contacts of staff with participants. Unscheduled contacts will be recorded.

Masking

Participants will not be masked to intervention allocation. Blinded personnel will include the principal investigator, the co-investigators, and the personnel responsible for data management and analysis. Unblinded personnel will include the PD, RA, and MT instructors. Blinding the PD and RA would be impossible given that the PD will be responsible for monitoring of adverse events and the RA will be responsible for randomization and data collection. However, study participants will perform all assessments independently using a computer interface without RA involvement. Finally, although the MT instructors will be aware of the allocation status of the participants, they will remain masked to the study hypotheses.

Study Assessments – Visits 1, 2, and 3

Assessment Schedule

The full assessment package (Table 2) will take 90–120 minutes and will be completed at baseline (Visit 1), end-of-treatment (3 months; Visit 2) and follow-up (9 months; Visit 3).

Table 2.

Detailed Assessment Timeline

	Screening	Baseline	End-of- Treatment (3 mos)	Follow-Up (9 mos)
Consent form	X
HIPAA	X
Contact information	X
Eligibility check log	X
Cognitive impairment (MOCA)	X
Suicidal ideation (SDDS-PC)	X
Covariates
Demographics		X
Medical history		X
Heart Failure Knowledge (DHFKS)		X
Expectancy & treatment credibility		X
Social support (MSPSS)		X	X	X
Blood pressure		X	X	X
Anthropometric measures		X	X	X
Primary Outcome
Cognitive function (NIH-TB Fluid Cognition Battery)		X	X	X
Secondary Outcomes
Interoception (HTT, MAIA)		X	X	X
Heart Failure Self-Care (SCHFI v.7.2)		X	X	X
Cardiac vagal control (hf-HRV)		X	X	X
Mood (HADS)		X	X	X
Health-Related Quality of Life (KCCQ-12)		X	X	X
Self-reported medication adherence (Voils)		X	X	X
Sleep quality (PSQI)		X	X	X
Exercise capacity (6MWT)		X	X	X
Biomarkers (B-type natriuretic peptide)		X	X	X

Note: 6MWT = 6-Minute Walk Test; DHFKS = Dutch Heart Failure Knowledge Scale; HADS = Hospital Anxiety and Depression Scale; hfHRV = high frequency heart-rate variability; HTT = Heartbeat Tracking Task; KCCQ-12 = Kansas City Cardiomyopathy Questionnaire-12; MAIA = Multidimensional Assessment of Interoceptive Awareness; MOCA = Montreal Cognitive Assessment; MSPSS = Multidimensional Scale of Perceived Social Support; NIH-TB = NIH Toolbox; PSQI = Pittsburgh Sleep Quality Index; SCHFI v.7.2 = Revised Self-Care of Heart Failure Index; SDDS-PC = Symptom Driven Diagnostic System for Primary Care; Voils = Voils DOSE-Nonadherence measure – Extent of Nonadherence domain

Primary Outcome – Cognitive Function (Fluid Cognition Composite Score)

Cognitive function will be measured with the NIH Toolbox Cognition Battery,⁴⁰ which includes seven cognitive tests; two measure crystallized cognitive ability (i.e., vocabulary, reading) and five measure fluid cognitive functioning (i.e., working memory, memory, processing speed, attention, executive function). Specifically (Table 3): The Flanker test is a measure of executive function, attention, and inhibitory control. The Dimensional Change Card Sort test measures cognitive flexibility and attention. The Picture Sequence Memory test measures episodic memory. The List Sorting Working Memory Test measures working memory. The Pattern Comparison Processing Speed test measures processing speed. Performance on these measures correlates significantly with “gold standard” neuropsychological tests and they have high test-retest reliability, internal consistency, and convergent and divergent validity.^41–46 Each test lasts 3–5 minutes and will be completed using a study tablet. Participants will undergo a practice trial prior to the administration of each cognitive test.

Table 3.

NIH Toolbox Cognition Battery Subtests Measuring Fluid Cognitive Function

Subtest Name	Fluid Cognition Domain(s) Measured
Flanker	Executive function, attention, inhibitory control
Dimensional Change Card Sort	Cognitive flexibility, attention
Picture Sequence Memory	Episodic memory
List Sorting Working Memory	Working memory
Pattern Comparison Processing Speed	Processing speed

The primary outcome measure is the Fluid Cognition Composite score at end-of-treatment. This score is obtained by averaging the normalized score of each fluid cognition measure and then deriving scale scores based on this new distribution. Higher scores indicate higher levels of functioning. Scores of 85–115 indicate average fluid cognitive ability compared with others nationally. Scores 115–130 suggest above-average ability, while scores ≥130 suggest superior ability. Conversely, scores 70–85 suggest below-average ability, and a score ≤70 suggests significant impairment. Individual fluid cognition subtest scores will be used as secondary outcome measures.

Secondary Outcomes

Interoception.

The Heartbeat Tracking Task (HTT)⁴⁷ will be used as an objective measure of interoception. Participants are instructed to count the heartbeats they perceive within a set time interval while in a resting sitting position. Three trials of 25, 35, and 45 seconds, respectively, with a 30 second rest between each, will be completed. Participants will report their perceived heartbeat number to the RA while the actual number of heartbeats during the trial is measured with a Holter recorder. The HTT score is calculated according to the formula of Schandry⁴⁷ and ranges 0–1, with higher scores indicating greater interoception.

The Multidimensional Assessment of Interoceptive Awareness (MAIA) is a 32-item self-report questionnaire with eight subscales and will be used as a subjective measure of interoception.⁴⁸ Subscales include: Noticing, Not-Distracting, Not-Worrying, Attention Regulation, Emotional Awareness, Self-Regulation, Body Listening, and Trusting. Subscales are scored as the average of relevant responses, and response options for each item range from 0 (Never) to 5 (Always). This instrument has adequate-to-very good reliability, with Cronbach’s alphas 0.66–0.87 and >0.70 for five of the eight scales.⁴⁸ For each subscale, higher scores represent greater interoception.

HF self-care.

The Self-Care of Heart Failure Index (SCHFI v.7.2)⁴⁹ will be used to measure HF self-care. The SCHFI v7.2 is a 39-item validated self-report instrument with Self-Care Maintenance, Symptom Perception, and Self-Care Management scales. Across all items, response options range from 1 (Never) to 5 (Always). Scales are scored using a publicly available scoring algorithm resulting in standardized scores of 0–100 for each scale. Higher scores reflect better HF self-care in each domain.

Cardiac vagal control.

High frequency heart rate variability (hf-HRV) will be measured using the Digital SEER Light Holter ECG Recorder – General Electric Health and analyzed using GE CardioDay Holter software. Specifically, cardiac vagal control will be assessed as per established standards using the high-frequency band (represented by high frequency power in Ln msec) by averaging the high frequency values for 5-minute segments.^50–52

Additional Outcomes

Mood.

Depressed mood will be assessed using the Depression subscale of the Hospital Anxiety and Depression Scale (HADS).⁵³ This subscale contains seven items with response options from 0–3; the subscale score is computed as a sum of all items. Total subscale scores range from 0–21 with higher scores representing greater depressive symptoms.

Health-related quality of life (HRQL).

Health-related quality of life will be measured with the Kansas City Cardiomyopathy Questionnaire-12 (KCCQ-12),⁵⁴ a 12-item short form of the original KCCQ.⁵⁵ The KCCQ-12 assesses four domains of HF-specific HRQL, including Physical Limitation, Symptom Frequency, Quality of Life, and Social Limitation. The 12 items have 5- to 7-point Likert-type response options, and domain scores are scaled such that each domain has a possible range of 0–100. A summary score is calculated as the average of the four domain scores, and higher scores represent better HRQL.

Self-reported medication adherence.

The Voils DOSE-Nonadherence measure – Extent of Nonadherence domain⁵⁶ will be used to measure self-reported medication adherence. This validated measure includes three items assessing 7-day frequency of missed medication doses from 1 (None of the Time) to 5 (Every Time). The domain is scored by taking the average of the three items; possible scores range from 1–5 and higher scores represent greater non-adherence.

Sleep quality.

Sleep quality will be assessed with the Pittsburgh Sleep Quality Index (PSQI),⁵⁷ a 19-item self-report questionnaire. The PSQI provides seven component scores: Subjective Sleep Quality, Sleep Latency, Sleep Duration, Habitual Sleep Efficiency, Sleep Disturbances, Use of Sleeping Medication, and Daytime Dysfunction, each with varying response formats (e.g., self-reported hours of sleep, Likert-style). Responses are converted to quantitative values via publicly available scoring guide with possible component scores ranging from 0–3. The seven component scores are summed to produce a global score ranging from 0 (better sleep) to 21 (worse sleep).

Exercise capacity.

The 6-Minute Walk Test (6MWT)⁵⁸ will be used as a functional measure of exercise capacity. The 6MWT measures the distance a patient can walk on a level course in six minutes, and it correlates with peak oxygen uptake⁵⁹ and predicted survival in HF drug trials.⁶⁰ Participants are instructed to walk for six minutes, “covering as much ground as they can in the allotted time.”⁵⁸ A higher value (in meters) indicates better exercise capacity.

B-type natriuretic peptide (BNP).

BNP, a marker of HF disease severity, ⁶¹ will be collected via venipuncture performed by a trained phlebotomist. Specimens will be collected in a plastic EDTA tube, centrifuged, and separated plasma samples will be stored at −20°C until testing. Assays will be analyzed using the ADVIA Centaur BNP assay, which measures BNP concentrations up to 5000 pg/mL with a minimum detectable concentration of < 2.0 pg/mL and a high specificity for BNP.

Covariates

We will collect information on participants’ sociodemographic characteristics (age, sex, race/ethnicity, education, marital status, income, insurance status) and medical history including medications using standardized self-report forms. Additional covariates will include HF knowledge (Dutch Heart Failure Knowledge Scale),⁶² social support (Multidimensional Scale of Perceived Social Support⁶³), participant’s expectations regarding the intervention (expectancy and treatment credibility assessment⁶⁴), blood pressure (BP; Dinamap XL automated BP monitor), and anthropometrics (weight, height, waist-to-hip ratio). All anthropometrics will be measured according to standardized protocols.⁶⁵

Safety

MT is a behavioral, low-risk intervention; however, minimal risk for potential adverse events is present and described below.

Emotional Discomfort

Though rare, when emotional discomfort occurs during mindfulness practice, it is typically mild and of short duration. At each session, the mindfulness instructor will actively inquire whether the participant has any psychological discomfort or if any distress has arisen during the previous’ week individual practice. If a patient presents signs of severe psychological discomfort, they will be excused from participating in MT. The instructor will inform the study psychologist who will advise on next steps. Participants will also be instructed to contact the PD should any discomfort arise during the training. If at any time during the intervention sessions the participant reports significant psychological distress, if severe depressive symptoms are detected at the HADS scale completion, or if the participant endorses suicidal ideation during eligibility screening, study staff will immediately initiate an assessment of suicidality risk. Specifically, the study psychologist will evaluate the patient over the phone and instruct staff on next steps (e.g., refer participant to the Emergency Room).

Breach of Confidentiality

Loss of privacy or confidentiality is rare and its impact on participants is likely to be minimal. Confidentiality of study data will be maintained by numerically coding all data, disguising identifying information, and keeping all paper data in locked file drawers. Electronic data will be stored in password-protected data files accessible only to authorized project personnel. Data containing identifiers (e.g., name, address, phone number) will be stored on password-protected secure servers separately from de-identified study data (e.g., surveys). All information obtained from participants will be accessible only to research staff. Audio recordings of MT classes will be reviewed promptly to ensure treatment fidelity and then destroyed. Instructions for maintaining privacy and confidentiality will be given to participants at the start of the first mindfulness class.

Venipuncture Side Effects.

Blood samples will be taken by venipuncture. Participants may experience mild discomfort associated with blood sample collection, including pain during blood draw and subsequent bruising. In very rare circumstances participants may experience fainting or local infection. These risks will be minimized by having an experienced certified phlebotomist performing the venipuncture.

Statistical Analyses

Prior to analyses, baseline characteristics will be summarized across the aggregate sample and between groups (MT+EUC vs. EUC). The distribution of primary and secondary study outcomes will be assessed and transformed as necessary (e.g., Log transform towards normality). Potential distributions for the outcome variables include normal and zero-inflated distributions.

All analyses described below will be conducted with the intention-to-treat (ITT) approach. For each aim, primary analyses will be ‘as randomized’ adjusting for baseline cognitive function (aim 1), interoception scores (aim 2), vagal control (aim 3) and stratification variables (gender, age).

Cognitive Function (Aim 1)

For our primary analysis, we will estimate effects of MT on the Fluid Cognition Composite score using a generalized linear model. Namely, the composite cognition score at end-of-treatment will be regressed on treatment group (MT+EUC vs. EUC), baseline cognition score and stratification variables (gender, age). Next, as a secondary analysis, we will use a mediation model based on a product of coefficients approach with bootstrapped standard errors (10,000 bootstrapped samples) to examine whether changes in cognition from baseline to end-of-treatment mediate treatment effects on self-care at follow-up. We will also explore the impact of MT on each separate cognitive domain of interest (memory, executive function, and attention) using a series of generalized linear models.

Interoception (Aim 2)

We will assess effects of MT on interoception (primary analysis) and whether changes in interoception at end-of-treatment will mediate improvements in self-care at follow-up (secondary analysis). Analyses will follow a similar plan as in Aim 1: generalized linear model to examine associations between MT and interoception at end-of-treatment; a series of mediation models based on a product of coefficients approach with bootstrapped standard errors (to examine path coefficients between MT, changes in interoception from baseline to end-of-treatment, and self-care at follow-up). Finally, we will examine associations between MT and additional outcomes of interest (mood, HRQL, biomarkers) using a similar analytic approach.

For both aims 1 and 2, self-care at follow-up will be analyzed with scores from each SCHFI v.7.2 scale in separate models. As these are secondary analyses, no alpha correction will be made for multiple comparisons.

Vagal Control (Aim 3)

We will estimate effects of MT on vagal control and whether changes in vagal control mediate improvements in cognitive performance at follow-up. We will focus specifically on the high-frequency band as an index of cardiovagal (parasympathetic) control.⁶⁶ For the primary analysis, we will assess associations with the Fluid Cognition Composite score; as secondary analyses, we will also assess the associations with each cognitive domain of interest (memory, executive function, and attention). Analyses will be similar to those described for Aims 1 and 2: a series of generalized linear models to examine effects of MT on vagal control at end-of-treatment (controlling for baseline); mediation models with a product of coefficients approach with bootstrapped standard errors.

Exploratory Analysis

Hospital HF admission rates will be calculated at follow-up and compared between groups using t-tests or non-parametric tests as appropriate.

Sensitivity Analyses

To increase rigor, we will conduct sensitivity analyses adjusting for time-varying covariates (e.g., unplanned contacts with study staff, comorbidity score, social support). We will also explore whether effects of MT on primary and secondary outcomes are moderated by number of sessions completed and duration of individual MT practice. Namely, we will include main effects of these moderators in the planned generalized linear models of primary and secondary outcomes, as well as interactions between the moderators and treatment group (MT+EUC vs. EUC).

Sample Size

The sample size (n=176) was calculated using G*Power and MPlus Monte Carlo estimation. Effect sizes from preliminary studies showed effects of MT on cognitive outcomes, including neurocognitive batteries, within the medium range (f²>.14) across various samples. Further, a recent meta-analysis from our group (under review) has shown effects of stress reduction interventions on vagal control with d=.23-.41. Assuming a two-sided alpha of .05 and medium-sized effects of MT, a total sample of n=158 would yield >90% to detect between-group differences in the primary and secondary outcomes at end-of-treatment. A series of Model Monte Carlo simulations were done to ensure adequate power for the mediation aims. Given indirect effects in the small-medium range, a total sample size of n=158 would yield >80% power to examine indirect effects of MT on HF self-care and HF biomarker outcomes. It should be noted that our pilot studies produced retention >95%,^18,20 thus we have conservatively inflated our sample size by 10% to allow for attrition, yielding a total sample of n=176 patients at baseline.

Discussion

Patients with comorbid HF and cognitive impairment display poor self-care and experience worse clinical prognosis than their HF peers.⁶⁷ Thus, identifying effective interventions to promote self-care in HF patients with cognitive impairment is an important step in improving HF outcomes. This RCT will expand upon our preliminary work which has supported MT interventions for promoting cognitive function in older adults at-risk for dementia and self-care in patients with HF.^18–20 Thus, the current RCT is a well-justified extension of these pilot studies and holds promise for identifying phone-based MT as an accessible intervention for promoting cognitive function and HF self-care in patients with comorbid HF and cognitive impairment.

The current study is a direct extension of our pilot single-arm, pre-posttest design study which showed encouraging results in support of MT as an intervention to promote self-care in patients with HF.²⁰ Here, we have increased rigor by adding a control group, increasing racial/ethnic diversity of groups (e.g., Spanish-speaking individuals), and ensuring a sufficiently large sample to limit type II error. Further innovations include the study of interoception as a pathway linking MT and improved self-care, and the investigation of vagal control as a mechanism whereby MT can improve cognitive function.

An additional strength of this study is the use of a phone-delivered MT. The adaptation of traditional MBSR to a phone-based protocol improves accessibility and reduces patient burden, making it appropriate for older adults with chronic medical conditions. Furthermore, because in this age group (>65 years) only 60% of patients own a smart phone and 3/4 say they are not very confident with technology,⁶⁸ app- or internet-based approaches are less suitable. Thus, a phone-delivered protocol is both highly scalable and particularly well-suited for the current sample. Our phone-delivered MT has been tested in over 140 patients with different chronic conditions (cardiac defibrillators, chronic HF, HIV), as well as pregnant women at risk of pre-term delivery, and has displayed high feasibility and acceptability and positive effects on a number of psychological outcomes.^34–37 While our modifications of the MBSR protocol pre-dated recently published recommendations for adaptations of MT interventions for specific populations,^69,70 these recommendations will be considered should future adaptations of our protocol be necessary.

In sum, phone-delivered MT has shown promise as an accessible, low-cost/low-tech intervention in HF patients with cognitive impairment.²⁰ By targeting cognitive function, this study addresses an important barrier to self-care in these patients. Demonstrating that, by improving cognitive performance and interoception, MT promotes self-care in patients with comorbid HF and cognitive impairment will pave the way to the integration of MT into rehabilitation programs to improve clinical outcomes in this high-risk and underserved group.